Hacker

Mohit Pradhan

@mohit230723

Joined Oct 2026Student at jntuhIndia

Elsewhere

Projects

  • LUNA

    Graph HacksOct 2026

    AI agents are good at generating text but bad at coordinating and being checked. Coding agents start every task without reusable skills, split work inconsistently, and merge it by hope. Investigation tools give answers you can’t trace back to the data. LUNA OS puts the plan, the skills, the evidence, and every decision into a FalkorDB graph. Coding agents connect through MCP: LUNA assigns skills from a skill graph to waiting sub-agent slots, checks that the pieces fit together, and assembles the result. On the website, you give LUNA documents or a database. It builds a context graph and answers questions by traversing it, then shows the exact path it walked, with each claim tied to source evidence. Who it’s for: Developers using coding agents, who want repeatable, inspectable multi-agent builds with skills they can add. Analysts and investigators, such as fraud, compliance, and due-diligence teams, who need to see connections between people, accounts, and events and trust where an answer came from.